Contributions of Loss of Lung Recoil and of

Enhanced Airways Collapsibility to the Airflow Obstruction

of Chronic Bronchitis and Emphysema

D. G. LEAvER, A. E. TATrERSFIELD, and N. B. PRIDE

From the Department of Medicine, Royal Postgraduate Medical School,Hammersmith Hospital, London, W.12. England

A B S T R A C T We investigated the contributions of in-trinsic disease of the airways, loss of lung recoil andenhanced airway collapsibility to the airflow obstruc-tion of 17 patients with chronic bronchitis and emphy-sema. Airways conductance at low flow (G.,), maxi-mumexpiratory flow (VE. MAX) and static lung recoilpressure [P.,t(L)] were measured at different lung vol-umes, and conductance-static recoil pressure and maxi-mumflow-static recoil pressure curves constructed. Lowvalues of AG.w/APa (L) and AVc, max/AP.t (L) wereattributed to intrinsic airways disease. Airway collapsi-bilitv was assessed by comparing G.w with upstreamconductance on forced expiration and by the interceptof the maximum flow-static recoil curve on the staticrecoil pressure axis (Ptm.).

All patients had reduced Gaw at all volumes but inseven GaGw/APatP(L) was normal. On forced expiration,maximum flow in all patients was reduced more thancould be accounted for by loss of lung recoil. AVE, MAX/AP.1 (L) was reduced in the patients in whonm AG.,/P,. (L) was low. In contrast AVE, MAX/AP,t (L) was nor-mal in three and only slightly reduced in another threeof the seven patients with normal AGaw/APat (L). Inthese patients Gaw greatly exceeded upstream conduc-tance and Ptm' was increased.

Weconclude that loss of lung recoil could account for,he reduction in resting airways dimensions in 7 of the17 patients. Enhanced airway collapsibility commonlycontributed to reduction in maximum flow. In threepatients the airflow obstruction could be entirely ac-counted for by loss of lung recoil and enhanced airwaycollapsibility.

Received for publication 16 August 1972 and in revisedform 26 February 1973.

INTRODUCTION

Pathological change in the airway wall and lumenleading to narrowing and occlusion is commonly con-sidered to be the most important abnormality in chronicairvays obstruction. Nevertheless, it has been recog-nized for many years that in some patients loss of lungrecoil and abnormal dynamic narrowing of airways mustcontribute to the expiratory difficulty (1, 2). Recentstudies have shown that lung recoil pressure is an im-portant determinant of the forces distending the airwaysduring quiet breathing (3, 4), and of the driving pres-sure for maximum flow on forced expiration (5, 6); inaddition, it has been shown that enhanced collapsibilityof the airways will reduce maximum flow (6). In thispaper we attempt to quantify the contributions of patho-logical change in the airway or its lumen of loss of lungrecoil and of enhanced airway collapsibility to the air-flow obstruction of 17 patients with "chronic bronchitisand emphysema." Studies have been made during quietbreathing and during forced expiration, and the resultsindicate that loss of recoil pressure plays an importantrole in reducing airway dimensions at low flow and thatenhanced airway collapsibility contributes to the reduc-tion in maximum expiratory flow (VE, MAX).' A prelimi-nary report has already been published (7).

' Abbreviations used in this paper: FEV1, forced expira-tory volume in 1 s; FRC, functional residual capacity;G.m, total airways conductance; G., conductance of the air-ways in the S segment; CGus, conductance of the upstreamsegment; P., (L), static transpulmonary pressure; Ptm',critical transmural pressure of flow-limiting airways; SGasspecific airways conductance; TLC, total lung capacity;VE MAX, maximum expiratory flow.

The Journal of Clinical Investigation Volume 52 September 1973 2117-2128 2117

VE

VI

A/ j//

,/,

,,.-;"I',7 -

FIGURE 1 Schematic isovolume pressure flow curve in mid-vital capacity. Conductanice is represented graphically bythe slope of the interrupted lines. Because the pressure-flowrelationship is curvilinear, conductance varies with thealveolar pressure applied. A represents the conductance atzero flow [in practice obtained by measuring the pressure-flow slope between zero flow and an inspiratory flow (V,)of 0.5 liter/sI. B represents the conductance at the mini-mumalveolar pressure associated with maximum expiratoryflow (PA'). C represents the conductance at a high alveolarpressure. The reduction in conductance between B and Cis accounted for entirely by increase(d dynaamic compressionof large airways in the downstream segment, while con-ductance of the upstream segment is unchanged (see text).

METHODS

Airways are elastic structures whose dimensions varywith changes in lung volume and with the dyinamic forcesdeveloped during breathing. Both these factors can becontrolled by studying the relation betw-een alveolar pressureand flow over a range of pressures at a single lung volume(8). A typical isovolume pressure-flow curve is shown inFig. 1. On such a curve, total airways conductance (flow/alveolar pressure) is represented graphically by the slopejoining any individual pressure-flowf point to the origin, soeven at a single lung volume, conductance varies continu-ously with the alveolar pressure applied. In this complexsituation there are two conditions under which at leastsome of the airway dimensions are fixed. The first is duringbreath-holding, when gas flow is zero and alveolar pressureequals atmospheric pressure. The second is at high valuesof alveolar pressure, where there is a plateau of expiratoryflow at the maximum value and flow becomes independentof the driving pressure. Although on the plateau of an iso-volume pressure-flow curve, total airw ay conductance de-creases in parallel with increases in alveolar pressure, theintrabronchial pressure measurements of Macklem and Wil-son (9) have shown that over the middle of the vitalcapacity in normal men this decrease is accounted for en-tirely by progressive dynamic compression of the largestintrathoracic airways. Measuring VE, MAX therefore not onlyassesses the maximum dynamic performance of the lungsbut can also provide information about the dimensions ofthe smaller airways (upstream segment of Mead, Turner,Macklem, and Little (5), see below) during forced expi-ration.

In the present patients we have examined these two pointson the isovolume pressure-flow curve by measuring con-ductance close to zero flow and VE, MAx at a variety of lung

volumes. As shown in Fig. 1, between these two poiInts onthe isovolume pressure-flow curve the pressure-flow rela-tion is curved. Although this is usually the relevant partof the curve duiring tidal breathingr, it is much more diffi-cult to analyse and we make no attempt to do so in thepresent paper.

Static dimensions of the airwcays: the conductance-staticrecoil plot. Our aim in this study was to distinguishchanges in airway size that are caused by reduction inairway-distending forces from those that reflect a changein the airway wall or lumen itself. We use the term in-trinsic disease of the airways to describe the situation whereairway narrowing has to be ascribed to changes in theairway wall or lumen.

In intact man the relationship between airway dimensionsand airway distending pressure can be approximated byplotting total airways conductance (Ga.), measured at lowflows, against static transpulmonary pressure [P.t(L)]. Con-ductance provides a relatively sensitive measurement of theoverall diameters of the tracheobronchial tree, because forlaminar flow, con(luctanice is proportional to the fourthpower of the radius of the airways. There is empiricalevidence that P*, (L) is a useful indicator of the forcesdistendinig the airwx-ays during breath-holding or breathingat low flows. In man, when the relation between P., (L)and volume is altered by strapping the chest wall, airwayconductance remains closely related to P.t(L) (3, 10).Hyatt and Flath (11) found that in dogs there wvere onlysmall differences between the diameters of intrapulmonaryairways when studiedl at a giveni Pst(L) in the intact lungand at the same transbronchial pressure after the lungparenchyma had been dissected away. Although these studiesindicate that P.t(L) is closely related to airway dimensions,this does not necessarily mean that peribronchial pressure isidentical with pleural surface pressure. A recent theoreticalanalysis by Mead, Takishima, and Leith (4) suggests thatthe forces (listending the bronchi would be exactly equiva-lent to Ps,(L) only if the specific compliance of alveoli andbronchi were equal and if there x-ere no tension in thelimiting membrane of the bronchus. \We know of no directcomparisons of specific compliance of bronchi and alveoli inhuman lungs; it has been su-gested that in dog lungs spe-cific compliance of bronchi may be less than that of alveoli(4), althoughl other recent experiments have not shown

NormalGaw

/KLoss of lung recoil}// Reduced bronchial

// ' distensibility"

Pst (L)

FIGURE 2 Schematic conductance-static recoil plot. Lossof lung recoil does not affect the conductance-static recoilslope. A reduced slope indicates a reduction in the overalldistensibility of the tracheobronchial tree.

2118 D. G. Leaver, A. E. Tattersfield, and N. B. Pride

a.

normalD

/loss of lung recoil

/ decreased Gus

//VE max

p t (L)

b. X

/ lopes/from origin

/X/< - Gus

pSt (L)

VE MO

C. /

I

-t' P6t (L)

FIGURE 3 Schematic maximum flow-static recoil plots. a. Loss of lung recoil does not affectthe maximum flow-static recoil slope. A decrease in the maximum flow-static recoil slopeindicates a decrease in upstream conductance. b. Parallel displacement of maximum flowpoints (heavy continuous line) to higher values of static recoil pressure. In the analysis ofMead et al. (5) this change is also interpreted as a decrease in upstream conductance. Inter-rupted line indicates a normal maximum flow-static recoil curve. c. Graphical interpretationof the maximum flow-static recoil curve (heavy continuous line) shown in b using the analysisof Pride et al. (6). See text.

consistent differences (12). If specific compliance of thebronchi were less than that of the alveoli, then the forcesdistending the bronchi would be systematically greater thantranspulmonary pressure (4). In plotting conductanceagainst P. (L), we are using recoil pressure as an indi-cator of the forces distending the airways and not neces-sarily implying that Pa (L) is precisely equivalent to thetissue forces distending the airways.

The change in conductance with change in static recoilpressure (AGaw/AP,t (L)) is a function of the radial dis-tensibility of the bronchial tree and is analogous to themeasure of bronchial distensibility in an isolated bronchusobtained by plotting bronchial diameter versus transbron-chial pressure. AG.aw/APsg (L), however, represents thelumped distensibility of all the bronchial tree and will beaffected by changes in any of the bronchial generations. Areduction in AGaw/APst(L) may indicate either stiffnessof individual airways or loss or complete occlusion ofparallel airways (Fig. 2); this is exactly comparable to theway in which a reduced lung compliance may reflect stiff-ening of individual alveoli or loss of alveoli. With loss oflung recoil without any intrinsic abnormality of the airwaywall or lumen, AGGaW/AP.g (L) will be within the normalrange.

Dynamic performance of the Ilungs: the mnaximitum flow-static recoil plot. Mead et al. (5) have shown thatVE, MAX at any lung volume where there is a plateau offlow on the isovolume pressure-flow curve can be analysedin terms of the relationship between Pat(L) and the con-ductance of the airways upstream to equal pressure points(conductance of the upstream segment, G..)

VE, MAX= P.(L) - (G.U). (1)

In normal subjects from about 70% down to about 30%of the vital capacity equal pressure points become fixed on

forced expiration at lobar or segmental bronchi (9), andso in these circumstances the upstream segment correspondsapproximately to the intrapulmonary airways. At lowerlung volumes in normal subjects and throughout the vitalcapacity in some patients with chronic airflow obstruction(13) the equal pressure points are situated in more periph-eral airways.

Eq. 1 can be expressed graphically by plotting maximumflow against static recoil pressure (Fig. 3a). With loss oflung recoil pressure AVE, MAX/APat (L) will remain in the

normal range. A reduction in AVE, MAX/AP.t (L) indicatesa decrease in upstream conductance. In addition a paralleldisplacement of maximum flow points to higher static recoilpressures without any change in AVE, MAX/AP. (L) will beinterpreted in the analysis of Mead et al. as a decrease inupstream conductance, which is particularly marked at lowrecoil pressure (Fig. 3b).

Pride, Permutt, Riley, and Bromberger-Barnea (6) de-veloped a slightly different equation to predict VE, MAx atany lung volume with a plateau of flow on the isovolumepressure-flow curve:

(2)

where Ptm' is the critical transmural pressure (differencebetween lateral airway and extra-airway pressure) at whichthe flow-limiting airways narrow sufficiently to restrictflow, and G. is the conductance of the airways in the Ssegment during forced expiration. In this analysis the Ssegment comprises all the airways between the alveoli andthose which form the flow-limiting segment. Hence the Ssegment will be shorter than the upstream segment of Meadet al. (5) if Ptm. > 0 and longer than the upstream segmentif Pt.' <0. This equation distinguishes two different ab-normalities of the airways, a decrease in conductance ofthe smaller airways (G.) and an enhanced collapsibility offlow-limiting airways (shown by an increase in Ptm'),which in the analysis of Mead et al. would both be inter-preted as a decrease in upstream conductance. If G. andPtm' are unaffected by changes in lung volume, their valuescan be derived from a maximum flow-static recoil plot, G.being represented by the slope of the maximum flow-staticrecoil points themselves (AVE, MAX/AP.t (L)) and Ptm' bytheir intercept on the static recoil axis (Fig. 3c). In thisanalysis parallel displacement of maximum flow points tohigher recoil pressures would be interpreted as being dueto an increase in Pt.', implying enhanced collapsibility offlow-limiting airways (Fig. 3c).

Experimental procedure. In each subject we measuredtotal airways conductance, VE, MAx, and P,, (L) over aswide a range of lung volumes as was possible. Total air-ways conductance (Gaw) was measured in a variable pres-sure, constant volume body plethysmograph by the pantingtechnique (14). As discussed above, the aim was to measureconductance close to zero flow so that there would be no

dynamic effects on airway dimensions. Patients with severe

Airflow Obstruction of Chronic Bronchitis and Emphysema

VE,mOx

VIE, MAX (Ps t (L) Pt.') - (G.).

2119

x

000 - -

o o o- oooo o o-S I+ +II

en O 0(1 sC1 tr. in1 t_ - -_fU^ t0 r; -t 1o -l I'd 10 1.114 -tt 1s *f) It *- u: In C4 - esq rs - " 1 - N, - o

,C ' t r, t" Ce ,t '" ,C Co :: - - nc - 'o 't

6~. 6

+,_ tr, (S 4 M O s C+++++++++++++++++ +

E4 tS X 0 tl. tr, \0 0 X X - f, fl0, x0 o x0 u, t- _ 4 o t- _ t -O - - - O 5- - - - O C

0 5C C 6 : C CO

.n U

0~~~~~~~~~0

f,q 0

00-,

--toc~0

00'= c,N N N-.

-11 r- It o .oc.-

')~= N-4- 47 = -~ - - =-: =

vS

6 --:= -6=_,_--__---: 1- Z

000 I

tl - -t I, ,l' if; 'o X -, -t -S tr, -n -S t- r- V', "d - == 0 = 0 0 0 0 C O O _ O0 0 0 = 0j: j i - j j~ j 0 j i C j i i j -Z _ ji

tn Cz l, cl,C,N

c 00 lll -1t-- tNN-

C,C't 't 00 t- oo 't 00 C~-'-"IN~ ~~06 ; N1-~t-

C t0 t--t

.~I"

r- X0 -S

0 -=U- .R -S,

el

Ut, 11

_- 5

-ZC

= O

Cv t-4

-0

0; C-

-.e-I4 "I)-C 0 m a,0 a) ~

z : 7 n

2120 D. G. Leaver, A. E. Tattersfield, and N. B. Pride

-1

e

e

0>

Us

04

>U

UCUdU) -

C)

.-4cd

-Z7

Wo

0 ,0

UCZ>

Cd

CS

U)._

=d00

:>

o

E o

> C d

cd 0

"O Y

QC," - C

CU

U 'V_

0

> -o 11

_I 6 _

airflow obstruction show considerable looping of the ex-piratory pressure-flow curve during panting, so we measuredthe slope between zero flow and an inspiratory flow of 0.5liter/s, during panting at about 2 breaths/s. Dynamicchanges on inspiration at these flow rates are small. Anaverage of 15 measurements over a range of lung volumeswas made and a line of best fit drawn by eye through thepoints. Total lung capacity (TLC) was determined by ob-taining thoracic gas volume by panting against a closedshutter (15) and then adding the volume inspired as thepatient made a full inflation on opening the shutter. Themean of at least four measurements was taken. Functionalresidual capacity (FRC) and residual volume were derivedfrom separate records of tidal breathing and vital capacitymaneuvers. VE, MAX-volume curves were obtained with thepatient seated within a variable-volume body plethysmo-graph (16) and breathing out to the room via a Fleisch No.3 or 4 pneumotachograph. The box-spirometer system waspressure-compensated and the amplitude and phase of vol-ume signals showed no change with increasing frequency upto 6.5 cycle/s. The pressure drops across the pneumotacho-graph screens were linear up to flows of 400 liter/min (No.3) and 600 liter/min (No. 4). Expiratory flow was plottedagainst change in thoracic gas volume on a Tektronix 564storage oscilloscope (Tektronix, Inc., Beaverton, Ore.).A series of full vital capacity maneuvers was performedwith efforts varying up to maximum and the largest expira-tory flows at a given volume were recorded (17). Pa (L)was measured with an esophageal balloon containing 0.5 mlair (18) at frequent intervals in the vital capacity, volumebeing measured with the patients seated within the variable-volume body plethysmograph. A mean pressure-volumecurve was constructed for each patient from at least threeclosely reproducible curves; occasional curves with morepositive esophageal pressures were ignored. Static expiratorycompliance was taken as the slope of the expiratory pres-sure-volume curve between FRC and 500 ml above FRC.The forced expiratory volume in 1 s (FEV1), and vitalcapacity were measured with a timed spirometer.

From these measurements, conductance-static recoil andmaximum flow-static recoil plots were constructed as fol-lows: (a) Conductance-static recoil plot, conductance ateach lung volume was read off the conductance-volumecurve and plotted against static recoil pressure for thisvolume. In all subjects an attempt was made to measureboth inspiratory and expiratory pressure-volume curves, butin most subjects the expiratory pressure-volume curveswere more repeatable. There was relatively little hysteresisbetween inspiratory and expiratory pressure-volume curvesi. the patients and so expiratory pressure-volume curveswere used in all except one patient in whom esophagealspasm at full inflation prevented reliable expiratory curvesbeing obtained. (b) Maximum flow-static recoil plot, P,a(L)obtained on the expiratory pressure-volume curve wasplotted against the value of V'E, MAX found at the samevolume on the VE, MAx-volume curve. Flows close, to peakexpiratory flow were ignored, since these are effort-de-pendent. The slope of the maximum flow-static recoil points,AV19,MAX/AP.t(L), was estimated between 70% and 30% ofthe vital capacity to obtain G.; the intercept of this slopeon the pressure axis was used to estimate Ptm.,

All volumes were expressed at body temperature andpressure.

Suibjects studied. 10 normal subjects (9 men and 1woman, average age 34.5 yr) and 17 patients with chronicairflow obstruction (16 men and 1 woman, average age 61.4yr) were studied. Only two of the normal subjects werecigarette smokers and none had any history of serious respi-

0.24

0I 0.20 -

Eu

X 0.16

NiC)X 0.12 -

.2

ff0.08

_

T00

0.04 I

20 40 60 Ii 100%Predicted TLC

120 140 160

FIGURE 4 G.w (measured at low flow rates) plotted againstlung volume. To correct for differences in size of the sub-jects lung volume is expressed as percent of predicted TLCand for conductance flow rate is expressed as predictedTLC per second. The shaded area indicates the normalrange which extends upwards to higher levels of con-ductance. The horizontal interrupted line indicates a nor-mal value of conductance at FRC, corresponding to 0.5liter/s/cm H20 for a man with a TLC of 6 liter. The opencircles indicate the values of conductance when the patientswere asked to pant after a period of quiet breathing (seetext).

ratory disease. The patients all had chronic cough andexpectoration and were present or ex-smokers. They hadbeen under observation for at least 2 yr during which timethere had been little variation in airways obstruction, eitherspontaneously or after adrenergic bronchodilator drugs or(in most cases) after controlled trial of prednisolone. At-tempts were made to select patients who showed the clinicaland physiological features of either the "emphysematous"or the "bronchial" types of chronic airways obstruction(19).

RESULTS

Anthropometric data and measurements of static anddynamic lung volumes, static lung compliance, and air-ways conductance in the normal subjects and the patientswith airways obstruction are summarized in Table I.The patients were smaller than the normal subjects andthe predicted total lung capacity (TLC) (20) of thelargest normal subject was 192% of the TLC predictedfor the smallest patient. In the Table, flow rates havebeen expressed in liters per second, but in Figs. 4-8, anattempt has been made to correct for these great differ-ences in size by expressing all flow rates as predictedTLC per second. In normal children and adolescentsmaximum flow rates and conductance values have beenshown to be closely related to TLC (21 ).

FEV,., lung volumes, and static lung compliance.FEV1 was reduced, and functional residual capacity andresidual volume were increased in all the patients with

Airflow Obstruction of Chronic Bronchitis and Emphysema 2121

0.20

N

Eo

0

-)

.-T0

30

0.

0.16

0. 12

0.08

0.04

0

- ,.~~~~~~--or.0- :~-.-.-

- --. ~~~~~.0 -

0 4 8 12

Pst(L) cm H20

FIGURE 5 Gas (measured at low flow rates aas predicted TLC per second per centimeter Iagainst lung recoil pressure. The shaded arealower part of the normal range. The resultspatients are shown. In the three following figur8) only the patients who had a normal condtrecoil slope (indicated by the thick, continuouthose with a very abnormal conductance-static(indicated by the interrupted lines) are shown.with intermediate values of conductance-staticare indicated by thin, continuous lines.

airflowv obstruction (Table I). Values of TLlung compliance in the patients were more ATLC was often above 130% of the predicteeight patients had values of static lung complwere greater than those found in any ofsubjects.

Relation between total airways co nducta n

volume. In all patients the conductance-vcvere displaced to higher lung volumes than

mal subjects (Fig. 4). As a result specific aductance (SGa.w) at FRC (obtained by. cvalue of conductance at FRC by the valueliters) was abnormal in all the patients (1Iaddition, in 14 of the 17 patients the value ofat FRC was less than 0.4 liter/s/cm H20, xxally taken as the lower limit of normal. In ththree patients the absolute value of conduct

the normal range and the abnormality in specific con-ductance was entirely due to the increase in FRC.

Routine measurements of airways conductance inpatients with airflow obstruction are often made at avolume above FRC measured during quiet breathing,since lung volume tends to increase at the onset of thepanting maneuver. When the present patients wereasked to pant after breathing quietly in the normal tidalrange, the average lung volume at which conductancewsas measured was 0.4 liter greater than FRC and con-ductance averaged 0.08 liter/s/cm H20 greater than theValue at FRC. In Fig. 4 the open circles indicate theresults obtained when patients were asked to pant aftera normal tidal breathing pattern and without any de-liberate attempt to change lung volume. In the Table,values of conductance at the actual FRC have been re-corded.

In most patients the conductance-volunme slope had alarge intercept on the volume axis. As a result specificconductanice in the patients will not be independent ofthe volume at which it is measured, but will increase

____________ w-ith increase in lulg volume.16 20 Relationl betzceen. total airways conidutctanice (at loaw

flows) anid lung rccoil pressures. There was consider-able variation in the relation between conductance and

,nd expressed P.t(L) in the patients, with a nearly 10-fold differenceH20) plottedindicates the between the hiighest and lowest AGaw/.APat(L) slopesfrom all 17 (Table, Fig. 5). In seven of the patients these slopes fell

.es (6, 7, and w%ithin the range obtained in our normal subjects; inuctance-static these patients the abnormalities in conductance shown

reol sl)o pd in Fig. 4 could be attributed to a reduction in lung re-

The patients coil pressure at a given lung volume. In contrast, we as-recoil slope sume that the patients with low AGaw/APat (L) slopes

have intrinsic disease of the airway wall or lumen.\NVe were interested in whether there were differencesin the dy-lanaic performance of the lungs of patients

C and static with normal and abnormal AGaw/APat(L) slopes. Invariable, but subsequelnt figures therefore we have confined our analy-d value and sis to the seven patients wvith normal AGaw/GAP.t(L)liance which slopes and the six patients with the most abnormalthe normal slopes (Fig. 5), and have omitted those with intermedi-

ate abnormality. The mean FEV1 was 1.27 liter in thecc and lung patients with normal AGaw,/APut(L) slopes and 0.93lume points liter in the six patients with markedly abnormal slopes;in the nor- only one of the eight patients studied with an FEV1 of

tirways con- less than 1.0 liter had a normal slope. The mean ex-lividing the piratory static compliance in the patients with a normal

of FRC in -GawGP//\P (L) slope was 0.56 liter/cm H20, as com-Pable I). In pared to 0.37 liter/cm H20 in the group with a reducedconductance slope.Thich is usu- Relation between D7E, M4x and lung recoil pressure. Ine remaining all patients VE, MAX was reduced at any given value of

:ance was in lung recoil pressure. The abnormality of the maximum

2122 D. G. Leaver, A. E. Tattersfield, and N. B. Pride

0.8 1

N-

-J

V-

-0

a.-1

x0E

>s

0.7 I

0.6 1

0.5 [0.4 1

0.3 [0.2 1

0.1

0 2 4 6 8 10 12 14 16

Pst(L) cm H20

FIGURE 6 \TE, MAX (expressed as predicted TLC per second) plotted against lung recoil pres-sure. Lower part of normal range indicated by shaded area. Continuous lines indicate patientswith normal AG.W/AP,t (L), interrupted lines those with a marked reduction in AGaw/Pat (L).

flow-static recoil curves was greatest in the patients withlow AG.w/AP.t(L) values (Fig. 6).

G. in the normal subjects lay between 0.069 and 0.142predicted total lung capacity/s/cm H20; it was below thisnormal range in 13 of the 17 patients. Pt.' was abovethe upper limit of our normal range (+ 1.7 cm H20)in 13 patients so that commonly the maximum flow-staticrecoil curve in the patients showed abnormalities in bothslope (G.) and intercept (Ptm'). However the sevenpatients with normal AG8w/AP.t (L) values had rela-tively little reduction in G.. Three of these patients hadvalues of G. within the normal range and in three moreG, was between 0.062 and 0.068 predicted TLC/s/cmH20. In these seven patients the average value of G. was0.072 predicted TLC/s/cm H20 compared to mean valuesof 0.094 in the control subjects and 0.047 in the patientswith low AGaw/APat (L) values. The mean increase inPtm' was slightly less in the patients with a normalAGaw/APat (L) slope (+ 2.8 cm H20) than in those withreduced AGaw/APat (L) slopes (+ 4.0 cm H20).

Relationship between total airways conductance at lowflows and rE, MAX (Fig. 7). There were considerabledifferences betveen the two groups of patients. In gen-eral those with a low AGGaw/APat(L) slope had relativelygood VE, MAX in relation to the value of total conductanceat the same lung volume, while the reverse was found inthe patients with a normal AG.W/AP.t (L) slope.

DISCUSSIONThe striking feature of the results is that in 7 of the 17patients the observed reduction in conductance (mea-sured at low flows) could be accounted for by the lossof lung recoil pressure. In no patient, however, was the

loss of lung recoil pressure sufficient to account directlyfor the reduction in VsM. MAx. This difference betweenstatic and dynamic measurements is discussed below.

Limitations in the methods. Previous workers (3)who have related airway conductance to lung recoilpressure have obtained lung recoil pressure during oc-clusion of the shutter at the end of each period of pant-ing. In pilot studies we found that the scatter of suchdirectly obtained conductance-static recoil points wasat least as great as that of conductance-volume points; asour patients tolerated the repeated panting procedurebetter when they were required to do this without an

0.8

-i' 06

-J

X 0.4

E.>" 0. 2

0

I II:12II, I

;//i@ i

, rR -00.240.08 0.12 0.16

Gow(Predicted TLC/s per cm H2pFIGURE 7 Relation between Gaw (measured at low flows)and \VE, MAX. For both measurements flow rates are ex-pressed as predicted TLC per second. The shaded arearepresents the lower part of the normal range. Continuouslines indicate patients with normal AG.w/APat (L), inter-rupted lines those with a marked reduction in AG.w/AP. t(L).

Airflow Obstruction of Chronic Bronchitis and Emphysema 2123

esophageal balloon in place, we chose to construct ourconductance-static recoil curves by drawing lines ofbest fit through 12-18 conductance-volume points andthrough 3-5 expiratory pressure-volume curves in eachpatient. This difference in technique does not seem tohave systematically affected the results, since the AGaw/AP.t(L) slopes in our normal subjects are very similarto those previously reported (3). A possible source oferror is that we have compared our normal subjectswith patients who were consi(lerably older. W\e know ofno evidence as to whether the condtuctanice-static recoilcurve changes with age but wve founld no obvious (if-ference in the position or the slope of the curve betweenthe older and younger individuals in ouir nornmal series.Although the maximum flow-static recoil relationshipwas approximately linear in most of the normal sub-jects, in about a third of the patients it was curved sothat the ARE, MAX/APIt (L) slope declined as static re-coil pressure and lung volume were reduced. In suchcurves we estimated G. from the lower part of theAR,B.MAX/AP,t(L) slope between 50% and 30% of thevital capacity and Ptm. from the intercept of this partof the ARE, MAX/API t(L) slope, so as to derive the lowestpossible values of G. and Ptm'.

Significance of the conductance-static recoil plot. In7 of the 17 patients conductance-static recoil points werewithin the normal range so that there was no functionalevidence of intrinsic disease of the bronchi. This doesnot necessarily exclude the presence of anatomicalchanges in these airways, for the following reasons.(a) In diseased lungs conductance will tend to risewith increasing frequency of breathing (22) and willstrictly reflect static airway dimensions only if the mea-surement is made at low flow rates and at a very lowfrequency of breathing, whereas the present measure-ments were made at a frequency of about 2 breaths/s.However, values of conductance at 2 cycle/s in thepatients with normal conductance-static recoil relation-ships were greater than 0.2 liter/s/cm H20 at FRC, andin such patients changes in conductance between fre-quencies of 0.3 and 2 cycle/s are small (see Figs. 6and 9 in reference 22). Although we cannot excludelarge changes in conductance at frequencies less than0.3 cycle/s, it appears unlikelv that frequency dependenceof conductance played an important part in explainingthe findings in the patients with normal conductance-static recoil relationships, but it will have led us to un-derestimate the severity of the airway disease in the pa-tients with more abnormal values of conductance. (b)Even at a very low breathing frequency a total air-way conductance within the normal range cannot ex-

clude anlatotmiical changes in the airways, because themeasurement is relatively insensitive to abnormalitiesin the smaller airways (23), and furthermore reflects the

lumped effects of all the airway generations and somight result from the combination of narrowing insome airways and increased dimensions of other airways.

In the patients wvith reduced AG1v/AP1t (L) slopes,the major factor reducing static airway dimensions wasa reduction in bronclhial distensibility, but in additionthere was almost always some loss of recoil pressure.We were unable to distinguish whether the reduceddistensibility was due to stiffening of individual airwaysor due to a reduction in the number of parallel airwaysby destruction or occlusion of the lumilen.

The only previous study of the relationi between con-ductance and P,t (i,) in patients w%-ith chl-oniic airflowobstruction using the present method, appears to bethat of Butler, Caro. Alcala, and DuBois (3). In theirpatients "there was less than the normiial change in air-ways conductance from alterations of lunig elastic pres-sure" but they do not present any further (letails. Jonson(24) has developed a technique of plotting total lungresistance (obtained by the esophageal balloon metlhod)against lung recoil pressure during the course of an ex-piratory vital capacity maneuver. He found tlhree pa-tients. all with only slight impairment of FEV71, in whliolnthe relation between resistance and lung volune was

abniormal but the resistance-lung recoil pressure rela-tion was normal. In this method resistance was mea-sured at an expiratory flow of 1 liter/s, which will leadto significant dynamic effects on the airways, especiallyin the presence of airflow obstruction, so it is probablethat Jonson's technique will underestimate the numnber ofpatients with normal static airway dimensions at a stani-dard distending pressure. Other authors have attemptedto distinguish between the roles of loss of recoil pres-sure and of intrinsic airways disease on the basis of thevalue of airway resistance during quiet breathing orpanting at FRC (25. 26). Only three of our seven pa-tients with a normal conductance-static recoil curvewould have been picked out by this criterion. WNlhethera patient with a normal conductance-static recoil curvewill have an abnornmal conductance at FRC will dlependon wk-hether lung recoil pressure is normal or low at FRC.A better technique w-ould be to measure the value ofGas /P.t(L) at FRC (27), but there are advantages inmlaking this nmeasurement over a range of lung volunmes,since Fig. 5 shows that the distinction betweeni a nior-mal and an abnormal conductance-static recoil slope isleast obvious at low lung recoil pressures.

Significance of the maximum floew-static recoil plot.All the patients showed a greater reduction in VE, MAX

than could be accounted for directly by loss of lung re-

coil, so that all had decreased upstream condutctance[absolute value of Vt:, MAx divided by the value of P.t(l)(5)1 during forced expiration. This finding was ex-

pected in the patients with low AG,-/AP.e (L) slopes

2124 D. G. Leaver, A. E. Tattersfield, and N. B. Pride

0.07Gus/Low- U bus IG3wwU.7

Nx 0.06E

06 0.050.0

v0.02

0.0

0

0 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22

Gaw (Predicted TLC/s per cm H20)

FIGURE 8 Relation between G, (measured at low flows) and GU.. For both measurementsflow rates are expressed as predicted TLC per second. Continuous lines indicate patients withnormal AG.w/AP.t (L), interrupted lines those with a marked reduction in AG."/AP., (L).Diagonal lines from origin indicate the line of identity between Gaw and Gu, and the linewhere Gaw is twice G... Normal values of Gus were greater than 0.075 predicted TLC/s/cmH20 at all lung volumes.

but not in those with normal AG../AP5t (L) slopes.Previously Duffell, Marcus, and Ingram (28) have de-scribed normal maximum flow-static recoil relationshipsin about half of a series of patients with chronic airflowobstruction, and more recently Black, Hyatt, and Stubbs(29) have reported similar findings in patients in whomthe airflow obstruction was associated with al-antitrypsindeficiency. ai-antitrypsin levels were measured andfound to be normal in three of the present seven pa-tients with a normal AG../AP.t(L) slope while none ofthe remaining four patients showed any of the charac-teristic clinical features (early onset of disability, strongfamily history, or predominantly basal emphysema) thatsuggest al-antitrypsin deficiency. Although our failureto find patients with completely normal maximum flow-static recoil relationships could therefore be explained,if ai-antitrypsin deficiency was particularly associatedwith uncomplicated emphysema, we believe the explana-tion may be that we studied patients with more ad-vanced disease. The group of patients with normalmaximum flow-static recoil curves studied by Duffellet al. (28) had an average forced expiratory volumein 0.75 s of 1.44 liter, while two of the four patientsfound to have normal maximum flow-static recoil re-lationships by Black et al. (29) had a maximum breath-ing capacity within normal limits. Outside the presentseries we have observed patients with radiological evi-dence of emphysema and considerable reduction in P,t (L)(but without ai-antitrypsin deficiency) in whom Vu, MAX-Pat (L) curves were completely normal. In these sub-

jects, howvever, expiratory flow limitation and the disa-bility were relatively slight.

We have considered two possible explanations forour findings that the patients in whom conductance wasreduced in proportion to loss of static recoil showed adisproportionate reduction in 'VE,MAX. First, it may bethat an abnormality of smaller airways, present in bothstatic and dynamic conditions, is detected by measure-ments of maximum flow but not by measuring G.,.Second, all airway dimensions may truly be normal un-der static conditions but airways contributing to Gi,may show enhanced dynamic changes on forced expira-tion. In distinguishing these two possibilities it is use-ful to compare values of G.- with those for Gus on forcedexpiration at the same lung volume. In patients withsevere intrinsic airways disease, narrowing of airwaysof less than 2 mmdiameter is responsible for most ofthe reduction in G., (23). So in these patients, in-creased frictional losses in small airways will be thedominant factor in determining the values of both G.and Gu. and these values might be expected to be similar.In fact a close relationship between Gawand Gu. was foundin patients with low values of AG.w/AP.t(L), althoughGaw was somewhat higher than Gu. (Fig. 8). In partthis difference arises because Gaw may have been over-estimated by use of the panting maneuver in these pa-tients. In contrast, the patients with normal values ofAG.w/AP.t (L), although they had values of upstreamconductance similar to those found in the patients withlow AG.w/AP.t (L) slopes, had much higher values of

Airflow Obstruction of Chronic Bronchitis and Emphysema 2125

total conductance, which were usually two to threetimes the values of Gu. at the same lung volume (Fig.8). Sinice in the patients with normal AGaw/.APat(L)slopes there wvould be less tendency to overestimate con-ductance by making the measurement during rapid pant-ing, we postulate that these large differences arose eitherbecause increased dimensions of large airways offset theeffects of small-airwav niarrowing on G. or becausethere were enhanced dynamic changes in airways con-tributing to Gut on forced expiration. Although the ma-jor bronchii Imlav be slightly larger in patients with em-p)hysema than in normal subjects (30), the functionalsignificaince of this change is doubtful since the con-ductance of central airways (those wvith an internal(liameter greater than 2 mm) was definitely above thenormal range in only one of seven emphysematous lungsstudied at necropsy (23). Webelieve it is more likelythat the large differences between Gaw and Gus were dueto enhanced dynamic changes in the airways. Althoughsuch changes are most obvious in the major extrapul-monary airways, lateral pressures in all the intrapul-monary airways on expiration are less than alveolarpressure so that even the most peripheral airways tendto be narrower than during breath-holding at the samelung volume. If the effective compliance of upstreamairways is increased this could lead to enhanced dynamicchanges during forced expiration. Further analysis ofmaximumii flow-static recoil curves in terms of theAXM,MAX/APs, (L) slope over the range 70%'-30% ofthe vital capacity (G.) and of the intercept of this slopeon the static recoil axis (Ptm') supports the presence ofa dynamic effect in these patients. Gs was within or justbelow the normal range in six of the seven patients withnormal conductance-static recoil curves, while Pt.' wasincreased to a mean value of + 2.8 cm H20 compared tothe value of - 1.3 cm H120 in our ten normal subjects(Table I). Hence most of the abnormality in maximumflow-static recoil curves in these patients wvas due toparallel displacemenit of maximum flow points to highervalues of static recoil, a change which is compatible withenhanced collapsibility of flow-limiting airways (6).This analysis is open to criticism in that it assumes thatvalues of GJ and Ptm' are independent of lung volumeso that the relation between maximum flow and staticrecoil can be described in terms of a slope and intercept.Although in some of the present patients this relation-slhip was notably curvilinear, this was not the case inthose patients who had a normal relation between con-ductance and Ps t (L) (Fig. 7). Other studies have shownexamples of nearly linear maximum flow-static recoilcurves with normal slopes but with an intercept at posi-tive values of Ps,t(L) in patients with a clinical diagno-sis of emphysema (28, 29, 31). A positive Pt.' impliesthat airways were limiting flow when lateral airway

pressure was greater than extra-airway pressure. Wecannot explain this finding; increases in Pt.' have previ-ously been described in asthma and have been attributedto increase in the active tension of bronchial smoothmuscle (6), but this is unlikely to explain the resultsin chronic airflow obstruction. Obviously there is a needfor a more direct assessment of dynamic airway changesthani is provided by analysis of maximum flow-staticrecoil curves; nevertheless both the large discrepanciesbetween Gaw and Gu5 and the tendency to an increase inP,,,' without reduction in G. are most easily explainedby enhanced airway collapsibility of airways at, or onthe alveolar side of, the flow-limiting segment. Severalother studies have indicated that dynamic factors areimportant in the airflow obstruction of emphvsema. Inthe middle of the vital capacity VE, MAX may be less thana fitth of the value of maximum inspiratory flow at thesanme lung volume (32). Comiipliance of lobar bronchi onforced expiration has been showin to be increased inemphysematous patients compared to normal subjects(33); this abnormalitity may be explained by atrophicchanges in the airways, whlich have been shown to in-volve lobar and segmental bronchi and several divisionsof the bronchi beyond segmental bronchi (34), implyingthat alterations in airway compliance in emphysemamay be equally extensive.

Previously the value of Ptm' in normal subjects hadbeen estimated as between - 5 and - 10 cm H20 (6).In the present normal subjects. the value of Ptmn1' deriveddirectly from maximum flow-static recoil curves, was- 0.8 cm H20, wlhich agrees closely with the value of- 1.3 cm H20 found by Permutt and Hyatt ([31] andpersonal communication). As Pt.' is so close to zero,the two analyses of VE, MAX set out in Eq. 1 and 2 abovewill not differ significantly in normal subjects.

Role of extrabronchial factors in c/ironic airflow ob-strutction.. In 1951 Dayman (2) attempted to define theroles of intrinsic disease of the airway and of loss ofairway-distending forces in patients with severe airflowobstruction. He suggested that in some forms of emphy-sema, "intrinsic obstruction of the airways may beslight or even absent." The present results confirm thatin some patients with chronic airflow obstruction thereis no evidence of intrinsic obstruction during breathingat low flow rates. In general these were not the mostseverely disabled patients although they often had largeincreases in static lung compliance. In these patientsmaximum flow was reduced out of proportion to the re-duction in P.t(L), implying there must have been in ad-dition either fixed intrinsic disease of the airways (whichfor some reason was not detected by measurements ofGaw) or abnormal dynamic changes in the airways. Wehave argued above that this reduction in maximum flowcould be explained by enhanced collapsibility of flow-

2126 D. G. Leaver, A. E. Tattersfield, and N. B. Pride

limiting airways or airways on the alveolar side of theflow-limiting segment. Although enhanced collapsibilityof extrapulmonary airways presumably indicates achange in the airway wall itself, this is not necessarilythe case for intrapulmonary airways where enhanced col-lapsibility might also be due to loss of airway supportby lung tissue. Hence in the patients with a normal rela-tion between conductance and P..(L) even the resultson forced expiration could be due to loss of extrabron-chial support as originally proposed by Dayman (2). Inmost of the severest cases of chronic airflow obstruction,however, there must be an abnormality in either theposition or the slope of the maximum flow-static recoilrelationship, since recoil pressures are rarely reducedbelow about one-third of the expected values, whilemaximum flows may be reduced to one-tenth or less ofexpected values. In advanced disease there is probablyboth intrinsic airways disease and loss of airway distend-ing forces. A reduced conductance at a standard P.t (L)was found in all the lungs studied at necropsy by Hogg,Macklem, and Thurlbeck (23), while all but one of thepatients in the present series had some reduction in lungrecoil pressure. Wesuspect that loss of airway-distendingforces commonly contributes to chronic airflow obstruc-tion but is rarely the sole cause of the severest disability.

Note added in proof. Since this paper was submitted, ananalysis of the relation between total pulmonary conductance(obtained by a similar method to that used by Jonson [24] )and static lung recoil pressure has been published by H. J. H.Colebatch, K. E. Finucane, and M. M. Smith. (1973. Pulmo-nary conductance and elastic recoil relationships in asthmaand emphysema. J. Appl. Physiol., 34: 143). These authorsfound that in patients with emphysema but without coughand sputum the relationship between expiratory conductanceand lung recoil pressure during a single slow expirationwas similar to that in normal subjects, and that in normalsubjects the slope of this relationship did not change withadvancing age. These results support our argument that thenormal conductance-static recoil relationship found in someof our patients is unlikely to be explained by the respiratoryfrequency at which we measured airways conductance.

ACKNOWLEDGMENTSThis work was supported in part by grants from the Na-tional Fund for Research into Crippling Diseases, theWates Foundation, and the Tobacco Research Council of theUnited Kingdom.

REFERENCES1. Christie, R. V. 1934. The elastic properties of the

emphysematous lung and their clinical significance. J.Clin. Invest. 13: 295.

2. Dayman, H. 1951. Mechanics of airflow in health andemphysema. J. Clin. Invest. 30: 1175.

3. Butler, J., C. G. Caro, R. Alcala, and A. B. DuBois.1960. Physiological factors affecting airway resistancein normal subjects and in patients with obstructive respi-ratory disease. J. Clin. Inzvest. 39: 584.

4. Mead, J., T. Takishima, and D. Leith. 1970. Stress dis-tribution in lungs: a model of pulmonary elasticity. J.Appi. Physiol. 28: 596.

5. Mead, J., J. M. Turner, P. T. Macklem, and J. B.Little. 1967. Significance of the relationship betweenlung recoil and maximal expiratory flow. J. Appl.Physiol. 22: 95.

6. Pride, N. B., S. Permutt, R. L. Riley, and B. Brom-berger-Barnea. 1967. Determinants of maximal expira-tory flow from the lungs. J. Appl. Physiol. 23: 646.

7. Leaver, D. G., A. E. Tattersfield, and N. B. Pride.1971. Factors contributing to airflow obstruction inpatients with chronic bronchitis and emphysema. Bull.Physio-Pathol. Respir. 7: 479.

8. Fry, D. L., and R. E. Hyatt. 1960. Pulmonary mechan-ics: a unified analysis of the relationship between pres-sure, volume and gas flow in the lungs of normal anddiseased subjects. Am. J. Med. 29: 672.

9. Macklemn, P. T., and N. J. Wilson. 1965. Measurementof intrabronchial pressure in man. J. Appl. Physiol. 20:653.

10. Stubbs, S. E., and R. E. Hyatt. 1972. Effects of in-creased lung recoil pressure on maximal expiratoryflow in normal subjects. J. Appl. Physiol. 32: 325.

11. Hyatt, R. E., and R. E. Flath. 1966. Influence of lungparenchyma on pressure-diameter behaviour of dogbronchi. J. Appl. Physiol. 21: 1448.

12. Hughes, J. M. B., F. G. Hoppin, Jr., and J. Mead. 1972.Effect of lung inflation on bronchial length and di-ameter in excised lungs. J. Appl. Physiol. 32: 25.

13. Macklem, P. T., R. G. Fraser, and W. G. Brown. 1965.Bronchial pressure measurements in emphysema andbronchitis. J. Clint. Invest. 44: 897.

14. DuBois, A. B., S. Y. Botelho, and J. H. Comroe, Jr.1956. A new method for measuring airway resistancein man using a body plethysmograph: values in normalsubjects and in patients with respiratory disease. J.Clin. Invest. 35: 327.

15. DuBois, A. B., S. Y. Botelho, G. N. Bedell, R. Mar-shall, and J. H. Comroe, Jr. 1956. A rapid plethysmo-graphic method for measuring thoracic gas volume: acomparison with a nitrogen washout method for mea-suring functional residual capacity in normal subj ects.J. Clin. Invest. 35: 322.

16. Mead, J. 1960. Volume displacement body plethysmo-graph for respiratory measurements in human subjects.J. Appl. Physiol. 15: 736.

17. Hyatt, R. E., D. P. Schilder, and D. L. Fry. 1958. Re-lationship between maximum expiratory flow and de-gree of lung inflation. J. Appl. Physiol. 13: 331.

18. Milic-Emili, J., J. Mead, J. M. Turner, and E. M.Glauser. 1964. Improved technique for estimating pleuralpressure from esophageal balloons. J. Appl. Physiol. 19:207.

19. Burrows, B., C. M. Fletcher, B. E. Heard, N. L. Jones,and J. S. Wootliff. 1966. The emphysematous and bron-chial types of chronic airways obstruction. Lancet. 1:830.

20. Cotes, J. E. 1968. Ltng function. Blackwell ScientificPublications Ltd., Oxford. 376, 380.

21. Zapletal, A., E. K. Motoyama, K. P. Van de WVoestijne,V. R. Hunt, and A. Bouhuys. 1969. Maximum expira-tory flow-volume curves and airway conductance inchildren and adolescents. J. Appi. Physiol. 26: 308.

22. Grimby, G., T. Takishima, W. Graham, P. Macklem,

Airflow Obstruction of Chronic Bronchitis and Emphysema 2127

and J. Mead. 1968. Frequency dependence of flow re-sistance in patients with obstructive lung disease. J.Clin. Invest. 47: 1455.

23. Hogg, J. C., P. T. Macklem, and W. M. Thurlbeck.1968. Site and nature of airway obstruction in chronicobstructive lung disease. N. Enigl. J. -Med. 278: 1355.

24. Jonson, B. 1970. Pulmonary mechanics in patients withpulmonary disease, studied with the flow regulatormethod. Scantd. J. Clin. Lab. Inivest. 25: 375.

25. Stevens, P. M., B. F. Orman, and G. Graves. 1969.Contributions of lung recoil and airway resistance toforced expiratory flow limitation. Am. Rev. Respir. Dis.100: 54.

26. Penman, R. W. B., R. P. O'Neill, and L. Begley. 1970.Lung elastic recoil and airway resistance as factorslimiting forced expiratory flow. A iner. Rev. Respir. Dis.101: 528.

27. Desmeules, M., R. Peslin, C. Saraiva, H. Uffholtz,and P. Sadoul. 1971. Proprietes statiques et dynamiquespulmonaires dans la bronchite chronique et l'emphyseme.Bull. Physio-pathol. Respir. 7: 395.

28. Duffell, G. M., J. H. Marcus, and R. H. Ingram, Jr.1970. Limitation of expiratory flow in chronic ob-

structive pulmonary disease. Relation of clinical char-acteristics, pathophysiological type, and mechanisms.Ann. Intern. Med. 72: 365.

29. Black, L. F., R. E. Hyatt, and S. E. Stubbs. 1972.Mechanism of expiratory airflow limitation in chronicobstructive pulmonary disease associated with a1-anti-trypsin deficiency. Am. Rev. Respir. Dis. 105: 891.

30. Fraser, R. G. 1961. Measurements of the calibre ofhuman bronchi in three phases of respiration by cine-bronchography. J. Can. Assoc. Radiol. 12: 102.

31. Permutt, S., and R. E. Hyatt. 1971. Analysis of iso-volume pressure-flow curves in normal subjects andpatients with airways obstruction. Bull. Physio-path.Respir. 7: 475.

32. Jordanoglou, J., and N. B. Pride. 1968. A comparisonof maximum inspiratory and expiratory flow in healthand in lung disease. Thorax. 23: 38.

33. Macklem, P. T., R. G. Fraser, and D. V. Bates. 1963.Bronchial pressures and dimensions in health and ob-structive airway disease. J. Appl. Physiol. 18: 699.

34. Wright, R. R. 1960. Bronchial atrophy and collapse inchronic obstructive pulmonary emphysema. Am. J.Pathol. 37: 63.

2128 D. G. Leaver, A. E. Tattersfield, and N. B. Pride